Designing a file transfer protocol that reaches the link throughput even when the round-trip-time is big and the percentage of dropped packets is also big is not very easy, but it is also not very difficult.
What would worry me more is how to be sure that the ISPs through which such a traffic passes, which is transported by a custom protocol, will not throttle or filter it.
If I download files over the Internet, it happens frequently for the speed to be less than the link would allow, but, at least for such speeds that do not exceed 1 Gb/s, the low speed never seems to be caused by TCP deficiencies, but always by some throttling that is done by some ISP, because the various lower speeds are always deterministic and associated with certain IP ranges.
If the Aspera protocol uses a large number of UDP destination ports, perhaps that confuses the ISP routers, which do not recognize this UDP traffic as belonging to a single connection that has a very high throughput, which could be higher than the ISP likes.
I don't think an equivalent alternative to Aspera will emerge due to IP encumbrances that will remain for longer than it will take QUIC to become an alternative to Aspera's FASP.
Agreed that direct legal action is low risk but until/unless IBM openly supports alternative implementations there is some unknown nonzero level of risk.
In my mind the natural incubator for such a QUIC capability is within the S3 user community. However since it seems useful and hasn’t happened yet I wonder what is going on.
What would worry me more is how to be sure that the ISPs through which such a traffic passes, which is transported by a custom protocol, will not throttle or filter it.
If I download files over the Internet, it happens frequently for the speed to be less than the link would allow, but, at least for such speeds that do not exceed 1 Gb/s, the low speed never seems to be caused by TCP deficiencies, but always by some throttling that is done by some ISP, because the various lower speeds are always deterministic and associated with certain IP ranges.
If the Aspera protocol uses a large number of UDP destination ports, perhaps that confuses the ISP routers, which do not recognize this UDP traffic as belonging to a single connection that has a very high throughput, which could be higher than the ISP likes.